Temperature Sensitivity of CO₂ and CH₄ Fluxes from Coarse Woody Debris in Northern Boreal Forests

Abstract

Carbon dioxide (CO₂) and methane (CH₄) are recognized as the main greenhouse gases causing climate warming. In forest ecosystems, the death of trees leads to the formation of coarse woody debris (CWD) that is one of the sources of greenhouse gas emissions due to wood decomposition. We quantified the CO₂ and CH₄ fluxes from CWD of larch (Larix gmelinii (Rupr.)) and birch (Betula tortuosa Ledeb.) collected in the northern boreal forests of Central Siberia. The CWD samples were incubated at +5, +15 and +25 °C. The CO₂ and CH₄ fluxes showed strong correlations with temperature, moisture, decomposition stage and the type of wood’s rot. The temperature coefficient Q₁₀ indicated higher temperature sensitivity of CO₂ flux within the temperature interval from +5 to +15 °C than from +15 to +25 °C. Methane flux had higher temperature sensitivity within the interval from +15 to +25 °C. It was found that, in boreal forests, CWD of early decay stage can serve as a source of methane to the atmosphere when air temperatures increased above +15 °C. Strong positive correlation between CH₄ production and CO₂ emission indicated a biological source and supported findings on aerobic origin of the main process contributing to the CH₄ flux from decomposing CWD.

1. Introduction

The increase in atmospheric greenhouse gas concentrations is one of the main reasons of the contemporary climate change [1], and carbon dioxide (CO₂) and methane (CH₄) play the most important role in acceleration of global warming. A significant amount of atmospheric carbon is sequestered in forest ecosystems: in the tree phytomass, dead plant residues and soil organic matter. According to the last estimates, annual carbon accumulation in Russian forests is 250 Mt, and more than 30,000 Mt of carbon is accumulated as wood biomass on the Russian territory [2]. During the growth and development of tree stands, the death of trees is leading to the formation of coarse woody debris (CWD). This is one of the least understood components of forest ecosystems [3]. Due to the high possibility of global carbon balance underestimation, the interest to CWD increased during the last decades [4].

Dead wood in the native boreal forests usually does not contribute much to greenhouse gases in the atmosphere due to the low rate of CWD decomposition [3]. However, in disturbed ecosystems (e.g., post-fire ecosystems, after the windfall, post-logging ecosystems and in the forests destroyed by pests), the amount of dead wood can increase significantly, thus turning into an important source of carbon dioxide emission [3,5,6,7,8]. Besides, the greenhouse gas emission from CWD can play a crucial role during spring and autumn, when the photosynthetic ability of plants is decreased [9]. This could be one of the reasons of observed seasonal fluctuations of atmospheric carbon in the Northern Hemisphere [9,10].

The field of our interest covers the permafrost region of Siberia. This region is of vital importance for the planetary processes since the strongest warming is observed in the high latitudes of the Earth [11]. More than 35% of the forest lands on permafrost in Russia are larch forests [12]. In Siberia, larch forms monodominant stands, which cover 84% and 93% of forested area in the northern boreal subzone and in the forest-tundra zone, respectively [13,14,15]. Birch forests are also widespread in the northern boreal subzone, especially in Central Siberia [16,17]. These northern forests are recognized as the most sensitive to the climate change [11,18] since they grow at temperatures which are currently limiting both biomass productivity and decomposition rate.

In the mature and old-growth larch forests of this region, CWD stock vary from 2 to 8 t ha⁻¹ [19], that is, 17–21% of total stand biomass in these ecosystems [20] or about 4% of ecosystem total carbon stock [21]. In the region under study, fire is the major disturbing factor, the average interval of fire events is about 80 years [22]. In postfire ecosystems, stock of CWD can increase by 60–100%, reaching 200% of tree stand biomass 10 years after the fire [19,20]. Carbon stock in this dead wood can comprise 6.5–8.0 tC ha⁻¹ in these postfire ecosystems [19].

A very significant increase of the surface air temperatures is already observed in the high latitudes [23], and the key question now is how strong different components of the northern forest ecosystems will react on this temperature rising? On the one hand, climate warming can lead to the increase of CWD stock due to the increment of forest fires and drought frequency in this region [24,25,26]. At the same time, higher temperatures can increase the rate of CWD decomposition [27]. Both of these developments might mean an enhancement of the flux of greenhouse gases from the pool of CWD from these forests [28].

One of the most common measure of temperature sensitivity mentioned in the scientific literature is the temperature coefficient Q₁₀ that shows how much the rate of the process is altered with temperature increase by 10 °C [29]. However, Q10 is a relative measure so that some other approaches should be involved to estimate the real values of the changed rate [30].

In our study, we decided to focus on CO₂ as well as on CH₄ fluxes, since both carbon forms are released from decomposing wood [31,32,33,34,35,36]. Carbon dioxide is a product of aerobic decomposer respiration. Methane production (methanogenesis) is mainly an anaerobic respiration [37]. Methanogens are found in all types of anaerobic environments. In addition, recent studies have shown biological and non-microbial methane formation under oxygenated conditions [37,38,39]. Despite the fact that methane production is possible in a wide range of conditions, not all produced CH₄ will be emitted to the atmosphere. Microbial aerobic and anaerobic CH₄ oxidation (CH₄ consumption) often takes place simultaneously, thereby counteracting the CH₄ production, and reducing the net CH₄ flux [37]. Methane-oxidizing microorganisms consume methane and metabolize it as a source of both carbon and reductant to drive cellular processes [40].

Recently, the measurement of the ratio of stable carbon isotopes have been widely applied to investigate sources and sinks of atmospheric CH₄ [41,42]. The natural abundance of stable carbon isotopes ¹²C and ¹³C comprise 98.89% and 1.11%, respectively, of the total carbon on Earth [43]. Measurements of these isotopes are expressed as the ratios to the more common ¹²C in a sample (¹³C/¹²C) and reported in the δ¹³C notation relative to the Vienna Pee Dee Belemnite (VPDB) standard in per mil (‰) [44]. Physiochemical and biological processes create distinctive carbon isotopic signatures in biogenic material that allow tracing its origin and identifying processes in the nature [45]. Mean carbon isotopic composition of atmospheric methane (δ¹³C-CH₄) is equal to −47‰ [46]. Microbially produced CO₂ and CH₄ are depleted in ¹³C in comparison with gases produced by chemical, photochemical or thermal decomposition of organic matter, because metabolic processes preferably use the lighter carbon isotope over the heavier isotope [47,48,49,50]. Molecules with the heavier isotope form stronger bonds and have higher activation energies, which makes them more difficult for utilization compared with molecules with lighter isotopes [45]. As a result, biogenic methane have δ¹³C equal to −60 ± 5‰, whereas biomass burning, coal mining and natural gas produce methane with δ¹³C varying from −24 ± 3 to −43 ± 7‰ [41]. Microbial oxidation of CH₄, both aerobic and anaerobic, is also associated with fractionation for C isotopes that leads to an enrichment of the residual CH₄ with the heavier isotope [51,52,53] and a depletion of ¹³C in the CO₂ pool [54].

Different field and laboratory incubation experiments have been performed, wherein the effects of temperature, moisture and decomposition stage on carbon loss from coarse woody debris were studied, such as the study on boreal black spruce in Manitoba [55], on Sitka spruce in Ireland [56], on Oriental oak and red pine in Korea [57,58], on CWD in Amazon tropical forests [59], and in temperate forests of China and Northern America [60,61]. Each of these studies only estimated the CO₂ flux from decomposing CWD and only one study focused on CH₄ emission from CWD decomposing by saprotrophic fungi [39]. However, to our knowledge, none of these studies have measured CO₂ and CH₄ emissions simultaneously, and the temperature sensitivity of CH₄ flux from CWD was not studied before.

In this study, we aimed to estimate possible changes in greenhouse gas emissions from CWD decomposition under the climate changes. The temperature sensitivity of CO₂ and CH₄ fluxes from decomposing dead wood of larch (Larix gmelinii (Rupr.) Rupr) and birch (Betula tortuosa Ledeb.), collected in the northern boreal forests of Central Siberia was studied in the laboratory incubation experiments. We quantified also the rates of methane production and methane oxidation in the decomposing wood through assessing the ratio of stable carbon isotopes.

2. Materials and Methods

The area of the study is Central Evenkia (64° N, 100° E) near the Tura settlement. The territory is hilly with gentle slopes within 120–600 m above the sea level, with pronounced erosive formation of valleys. The whole territory is underlain by continuous permafrost, and soils are Cryosols [62].

The climate is cold (continental) without a dry season according to Köppen-Geiger climate classification [63]. The mean annual temperature is −8.9 °C. The annual amplitude of temperature is 52 °C, and the growing degree-days above 10 °C is 1000 °C, while the frost-free period amounts to 70–80 days. Mean annual precipitation is 369 mm. Seasonal distribution of precipitation is approximately even. Snow cover thickness varies within 50–60 cm. The climate differs with the altitude due to the air mass inversions between the foot and summit of mountains [62].

In the studied area, larch (Larix gmelinii) and birch (Betula tortuosa) stands are predominant. Cowberry (Vaccinium vitis-idaea L.), bog bilberry (Vaccinium uliginosum L.), marsh tea (Ledum palustre L.) and typical boreal mosses (Pleurozium schreberi (Brid.) Mitt., Hylocomium splendens (Hedw.) Bruch et al., Aulacomnium palustre (Hedw.) Schwagr.) are widespread under the tree canopy. In general, birch forests in Evenkia occupy a small area, about 5%, of the overall territory [13]. However, in Central Evenkia, in the Northern boreal subzone, near the Tura settlement, birch forests (Betula tortuosa) are widespread, especially on the elevated flat mountain plateau and occupy about 30–40% of the territory [17].

The forest floor in studied ecosystems is covered by laying logs of different stage of decomposition. Decay class of logs was defined in the field according to the system of [64], and three decay classes of CWD were defined based on visual and physical properties. The principle of division was based on wood density and the presence of bark and branches, as follows:

• DC I: Wood has not lost its solidity; log has bark; small to medium branches are present;
• DC II: Wood has lost some of its solidity; bark easily flakes from wood, but bark and larger branches are present on the log;
• DC III: Wood has lost its initial solidity; some bark and large branches can remain on the log.

Discs of larch and birch laying logs were collected each from the same area. Samples were taken at least from three logs of the same decay class for larch and birch. In total discs of 26 logs were sampled. Diameters of the sampled logs varied from 5.7 to 14.8 cm, representing the prevalent tree diameter within the territory of the study.

For larch logs at DC III, it was possible to identify the type of rot associated with the activity of ligninolytic or cellulolytic basidial fungi (i.e., white rot or brown rot, respectively). The type of rot was determined in the field by visual signs. Thus, for larch at DC III, we measured CO₂ and CH₄ emission for wood decomposed by white and brown rot fungi separately.

Experimental Design

The discs of larch and birch logs at different stages of decomposition were placed in the separate gas-tight plastic boxes of 1500 mL volume. Boxes were hermetically sealed. Immediately after placing the samples in the box and closing, we measured initial concentrations of CO₂ and CH₄ and the stable carbon isotope ratios (δ¹³C) in these gases in the box headspace using Picarro G2201-i cavity ring-down spectrometer (Picarro, Inc., Santa Clara, USA). We incubated the samples at three different temperatures. At the beginning of the experiment, samples were incubated at +5 °C. The following measurements of CO₂ and CH₄ concentrations were made after 3 h, 24 h, 3 and 6 days. After that, boxes with CWD samples were opened and aerated. The same procedure was afterwards repeated also for other temperature levels (+15 and +25 °C). Before each series of measurements boxes with CWD samples were allowed to equilibrate under specific temperature of the series for 24 h.

At the beginning of each measurement series, the boxes were weighted to trace water loss after aeration. At each measurement, weight loss did not exceed 0.5 g that was less than 5% of sample water content. Woody discs were weighted before experiment and after termination of the experiment they were oven dried at 80 °C until constant weight. Water content (%) was calculated as the ratio of the mass lost (difference between initial and dried weight), attributed to water initially present in the sample, to total mass of the fully dried wood:

W = (M₁ − M₂)/M₂ × 100,

(1)

where W is the water content in the sample (%), M₁ is the initial weight of the CWD sample (g), and M₂ is the dry weight of the CWD sample (g).

Bulk density of wood was measured for sectors of the woody disc by water displacement method [65]. Carbon and nitrogen contents and the stable carbon isotope ratios in the wood (δ¹³C_wood) were measured using an elemental analyzer (Vario Isotope Cube, Elementar Analysis Systems GmbH, Hanau, Germany) coupled with IRMS (IsoPrime100, Elementar Analysis Systems GmbH, Hanau, Germany). Stable carbon isotope ratios were expressed using the conventional δ notation as ‰ deviations from the Vienna Pee Dee belemnite (VPDB) standard [44,66]:

δ¹³C = ((R_sample/R_standard) − 1) × 1000,

(2)

where R = [¹³C]/[¹²C] for sample and standard, respectively.

The rate of the CO₂ and CH₄ emission (S_CO2/CH4) was calculated based on the change in concentration of these gases in the box headspace after 24 h of incubation at different temperatures using the following equation:

S_CO2/CH4 = ΔCO₂(CH₄)/Δt × (V_air × M_CO2(CH4))/22.41 × 273.15/T_air × 1/W_CWD

(3)

where S is CO₂ or CH₄ flux (µg CO₂ g⁻¹ h⁻¹), ΔCO₂(CH₄)/Δt is the CO₂ or CH₄ concentration increment (ppm) per time unit (h), V_air is the air volume (liter) in the chamber calculated as V_air = V_chamb − V_sample (V_chamb is the volume of the chamber, V_sample is the volume of CWD sample), M_CO2(CH4) is the molar mass of CO₂ (44 g mol⁻¹) or CH₄ (16 g mol⁻¹), T_air is the air temperature (K), 22.41 is the molar volume (L mol⁻¹) at the standard temperature (273.15 K) and pressure (1.013 bar), and W_CWD is the CWD sample dry weight (g) [58].

Since, as well, CH₄ oxidation as emission could be taking place, the relative CH₄ increase cannot be used as a direct measure for CH₄ production. For this reason, the isotopic signature of CH₄ was used. It was expected that, during CH₄ consumption, the lighter ¹²C-CH₄ isotope was preferred. At the same time, CH₄ production was expected to not fractionate.

From here on, we will call the net emission, based on the measured concentration change, the “CH₄ flux”. The actual amount of produced CH₄, which we base on studying the isotopic composition of CH₄, we will call the “CH₄ production”. The calculated difference between the CH₄ production and the net CH₄ flux, we will call the “CH₄ consumption”.

The rate of methane production was calculated based on the assumption that emitted amount of ¹³CH₄ reflected the rate of methane production due to isotopic fractionation during methane oxidation. This rate was calculated as following:

• The amount of evolved ¹³CH₄ (Δ¹³CH₄) was calculated as a difference between concentration of ¹³CH₄ measured at the time t and initial ¹³CH₄ concentration in the box headspace;
• Using Equation (2) (the main equation for δ¹³C), we calculated which amount of ¹²CH₄ ([¹²CH₄]) could be produced if supposing that there was no fractionation during methanogenesis and produced methane had δ¹³C similar to that in the substrate (wood):

  R_CH4 = [¹³CH₄]/[¹²CH₄] = R_standard × (δ¹³C − CH_{4_prod}/1000 + 1),

  (4)

  where δ¹³C-CH_{4_prod} is a stable carbon isotopes ratio in produced methane and δ¹³C-CH_{4_prod} = δ¹³C_wood.

The amount of ¹²CH₄ calculated from Equation (4) is the expected change in ¹²CH₄ concentration (Δ¹²CH₄) in case there would be no methane consumption in decomposing wood.

3.

The rate of methane production was calculated for the sum of Δ¹²CH₄ and Δ¹³CH₄ using Equation (3).

In these calculations we assumed that there was no fractionation of stable carbon isotopes during methane production in decomposing wood. The stable carbon isotope ratio in methane is reported to vary from −22‰ to as low as −120‰ [45], and the lowest δ¹³C values are related to microbially produced methane [67]. However, these values were received mainly for anaerobically produced CH₄. There is no data on the ratio of stable carbon isotopes in methane produced under aerobic conditions. Additionally, we do not know clearly which agents or processes produce methane during wood decomposition. Thus, we supposed that stable carbon isotope ratio in produced methane might be similar to that in the substrate (in the decomposing wood).

The volume of woody samples, which had been incubated for CO₂ and CH₄ flux measurements, varied from 119 to 581 cm³. As it was shown by Yoon et al. [58], the rate of respiration and methane fluxes from CWD calculated on the weight basis does not depend on the size of wood sample.

To analyze the dependence of CO₂ and CH₄ fluxes on temperature, we calculated relative and absolute measures of temperature sensitivity. We used the Q₁₀ coefficient to characterize the relative dependence of CO₂ and CH₄ fluxes on temperature [30]. This coefficient represents the factor by which the rate of biological or chemical processes increases with the temperature increment of 10 °C. This factor was calculated as follows:

Q₁₀ = (S₂/S₁) ^{[10/(T2 − T1)]}

(5)

where S₁ and S₂ are the rates of CWD respiration or methane fluxes at temperatures T₁ and T₂ [30].

The absolute temperature responses of CO₂ and CH₄ fluxes were calculated as the first derivative of a linear model connecting two rates measured at different temperatures. This model supposes monotonic increase or decrease of the rate within the studied temperature interval. The slope of the line showed how much the rate of studied fluxes had changed with temperature. Calculated temperature response showed changes of the studied fluxes with each degree of temperature.

To avoid confusion we used term “temperature sensitivity” for Q₁₀ values and “temperature response” for the absolute changes of CO₂ and CH₄ flux rates with respect to temperature.

Differences in the rate of CO₂ and CH₄ fluxes, and Q₁₀ values between species, decomposition classes, and temperature diapasons were analyzed using Mann–Whitney U test. Analysis of variance (ANOVA) and analysis of covariance (ANCOVA) were conducted to assess the overall importance of each influencing factor (tree species, wood density, temperature, and water content).

3. Results

3.1. Bulk Density, Water Content, and Chemical Composition of Studied CWD

Average bulk density of woody samples used in the incubation experiment decreased from DC I to DC III almost twice for larch and almost three times for birch CWD (

Table 1. The main characteristics of the coarse woody debris samples used in the experiment.

Decomposition Class	Bulk Density, g cm−3	Water Content, %	Carbon		Nitrogen, %	C:N
			%	δ13Cwood, ‰
Larch
DC I	0.392 (0.018)	41.8 (11.9)	47.6 (0.14)	−26.4 (1.48)	0.060 (0.008)	819 (108.0)
DC II	0.396 (0.019)	29.4 (10.2)	47.8 (0.25)	−24.8 (0.33)	0.092 (0.008)	564 (66.9)
DC III w	0.204 (0.031)	46.7 (25.5)	47.2 (0.19)	−25.3 (0.30)	0.267 (0.116)	201 (57.6)
DC III b	0.291 (0.014)	99.4 (48.0)	49.5 (0.54)	−26.6 (0.40)	0.235 (0.061)	229 (53.5)
Birch
DC I	0.449 (0.015)	117.6 (4.5)	47.5 (0.28)	−26.7 (0.43)	0.157 (0.041)	322 (69.1)
DC II	0.373 (0.020)	163.7 (38.9)	47.4 (0.19)	−27.2 (0.26)	0.109 (0.007)	436 (30.8)
DC III	0.159 (0.014)	449.9 (79.7)	47.2 (0.97)	−27.1 (0.17)	0.348 (0.067)	141 (22.9)

Values are the mean (±SE) of three-seven replicates. Index letters: w—for white rotted wood; b—for brown rotted wood.

). Water content of birch wood samples was three to five times higher than in the samples of larch CWD. Carbon content was similar for samples of all decomposition classes and for both tree species. Only larch wood at the DC III being decomposed by brown rot differed from the other samples by significantly (p < 0.05) higher content of carbon. This wood was noticeably depleted in ¹³C in comparison with wood decomposed by white rot and DC II. Shift of the carbon isotope ratios (δ¹³C) with decomposition stage was substantially lower for birch wood in comparison with larch wood. The nitrogen content gradually increased from DCI to DC III for wood of both tree species, resulting in lower C to N ratios at the late stage of decomposition. Birch wood at all decomposition classes contained significantly larger amounts of nitrogen than larch wood (p < 0.05) (Table 1).

3.2. CO₂ Flux from Decomposing CWD Samples

Curve shapes of cumulative carbon dioxide yield during 144 h incubation indicated saturation with time (Figures S1 and S2). The δ¹³C ratio in CO₂ decreased during incubation from ambient −12–−13‰ to −23–−33‰. The δ¹³C-CO2 equal to −23–−33‰ is the result of mixing: the produced CO₂ mixed with CO₂ of the air (with δ¹³C = −12–−13‰). In order to shift the ambient isotopic ratio to values that produced CO₂ it must have a lower isotopic signature than in the wood (−24–−26‰). It means that some isotopic fractionation took place during CO₂ production, which is a sign of biological decomposition of wood organic matter. The calculated rate of CO₂ emission during the first 24 h of incubation showed strong dependence on temperature (Figure 1).

At the advanced stage of decomposition (DC III), the CO₂ flux from larch CWD mostly depended on the fungi participating in the decomposition. Decay by brown rot fungi resulted in the release of more than twice a higher amount of CO₂ in comparison with wood decomposed by white rot.

Birch wood produced a higher amount of CO₂ than larch wood during incubation at temperatures +15 and +25 °C. The highest rate of CO₂ emission from birch wood was observed at the decomposition stage DC III.

3.3. Methane Fluxes from Decomposing CWD Samples

Measured flux of CH₄, calculated as methane concentration changes in the headspace of the experimental box during incubation, is the result of two contrary directed processes: methane production and methane consumption. It characterizes the rate of CH₄ flux during decomposition of CWD. In some cases, when during incubation the concentration of CH₄ decreased in the experimental volume, the rate of methane flux was negative indicating evident methane oxidation. However, even if CH₄ increased in the headspace during incubation, CH₄ uptake might still take place, but it was just not dominant.

The calculated rate of CH₄ exchange had no significant dependance on temperature; however, some trend of negative rates of this flux was observed for DC III both for larch and for birch wood (Figure 2a).

During exposition of larch CWD at +5 °C the average flux of CH₄ decreased strongly from DC I to DC II, and from DC II to DC III decomposed by brown rot (from 6.14 × 10⁻⁵ ± 3.36 × 10⁻⁵ µg CH₄ g⁻¹ h⁻¹ to 2.27 × 10⁻⁵ ± 0.69 × 10⁻⁵, and to 1.67 × 10⁻⁶ ±1.32 × 10⁻⁶ µg CH₄ g⁻¹ h⁻¹, respectively). During exposition at +15 °C the average rate of CH₄ flux did not change noticeably for DC I and DC II. CWD at DC III showed evident increase of CH₄ oxidizing (the average rate of CH₄ flux decreased to −7.81 × 10⁻⁶ ± 0.35 × 10⁻⁵ and −4.99 × 10⁻⁶ ± 0.45 × 10⁻⁵ µg CH₄ g⁻¹ h⁻¹ for wood decomposed by white rot and brown rot, respectively). Incubation at +25 °C showed an increasing CH₄ flux from DC I and DC II. Concerning DC III, the methane flux from CWD increased in some cases of white rot decomposition. In contrast, larch CWD decomposed by brown rot showed an increase of the methane oxidizing rate (the average rate of CH₄ flux decreased to −2.6 × 10⁻⁵ ± 0.51 × 10⁻⁵ µg CH₄ g⁻¹ h⁻¹).

Incubation of birch CWD demonstrated evident methane flux only for DC I. The average rate of this flux increased more than forty folds during incubation at +5 and +25 °C (from 2.01 × 10⁻⁶ ± 1.02 × 10⁻⁵ to 8.14 × 10⁻⁵ ± 6.53 × 10⁻⁵ µg CH₄ g⁻¹ h⁻¹, respectively). Birch wood at DC II and DC III showed methane oxidation during incubation at all temperatures (Figure 2).

During incubation, δ¹³C of CH₄ increased from initial −50–−54‰ up to +60.2–+74.0‰. CWD samples that emitted CH₄ showed a linear relationship between the increase of ¹²CH₄ and ¹³CH₄ concentrations in the experimental boxes (Figure 3). However, CWD from birch at DC II and from both tree species at DC III showed the decrease or no changes in the ¹²CH₄ concentration during incubation. However, for these samples, we observed the significant increase of δ¹³C of CH₄ due to evident rising of ¹³CH₄ concentration (Figure 3).

Based on the changes of the absolute ¹³CH₄ concentration in the headspace of the experimental boxes during CWD incubation, we calculated the possible rate of ¹²CH₄ production, supposing that carbon isotopic ratio in the produced methane should be similar to that in the substrate—decomposing wood (Table 1). The calculated rate of CH₄ production in all cases exceeded the measured rate of CH₄ flux and showed a significant dependance on temperature (Figure 2b). Based on the difference between rates of calculated methane production and measured methane flux, we assessed the rate of methane consumption (Figure 2c). The average rate of methane consumption increased from DC I to DC III for CWD of both studied tree species. Exception was DC II, which demonstrated the low rate of CH₄ consumption for both species at temperatures of +15 and +25 °C.

3.4. Temperature Sensitivity and Temperature Response of CO₂ and CH₄ Fluxes

To characterize the dependence of CO₂ and CH₄ fluxes on temperature we calculated coefficient Q₁₀ for the temperature intervals +5 to +15 °C and +15 to +25 °C. For the CO₂ fluxes from larch wood at early stages of decomposition (DC I and II), the average Q₁₀ values were larger for temperature interval from +5 to +15 °C (2.87 ± 0.74 for DC 1 and 3.51 ± 1.85 for DC II) than for +15 to +25 °C (2.33 ± 0.62 for DC I and 2.41 ± 0.97 for DC II). At DC III, the average Q₁₀ for CO₂ fluxes from CWD was almost similar within both temperature intervals (3.16 ± 1.62 and 3.11 ± 1.23). Temperature sensitivity was quite different for CWD decomposed by white and brown rot fungi. For CWD decomposed by white rot, Q₁₀ for the temperature interval +5 to +15 °C appeared to be lower than that for the interval +15 to +25 °C (1.79 ± 0.77 and 4.01 ± 2.45, respectively). On the contrary, CWD decomposed by brown rot indicated higher temperature sensitivity for the lower temperature interval (+5 to +15 °C) in comparison with +15 to +25 °C (the average Q₁₀ value was equal to 4.29 ± 2.47 and 2.78 ± 0.82, respectively). However, factorial ANOVA analysis indicated that differences for decomposition classes and temperature intervals mentioned above were not significant due to the wide range of temperature sensitivity of CO₂ flux (Figure S3b).

The average coefficient Q₁₀ for CO₂ emission from birch CWD exceeded 3.00 for DC I and DC III (3.17–3.33 and 3.16–3.69 for DC I and III, respectively) (Figure S3a). Temperature sensitivity of CO₂ flux was significantly lower (p < 0.05) for DC II, with Q₁₀ values of 2.84 ± 0.15 and 2.65 ± 0.23 for temperature intervals +5 to +15 °C and +15 to +25 °C, respectively. Differences in the temperature sensitivity of CO₂ flux between two studied temperature intervals were nonsignificant (p > 0.05).

Methane fluxes had higher temperature sensitivity (Q₁₀) in the temperature interval +15 to +25 °C than at the lower temperatures (+5 to +15 °C). Methane flux increased by 30–48% when temperature raised from +5 to +15 °C, while the increase was about four to eight-fold in case of the temperature increase from +15 to +25 °C. The average Q₁₀ for the methane production was 1.14–1.68 and 2.28–3.42 for temperature intervals +5 to +15 °C and +15 to +25 °C, respectively. Temperature sensitivity of methane consumption for wood at DC I was similar within both studied temperature intervals. For DC II and DC III, Q₁₀ for this flux increased from temperature intervals +5 to +15 °C to +15 to +25 °C (from 1.16–1.76 to 2.42–2.72, respectively) (Figure S3).

Decomposition class and temperature interval significantly influenced Q₁₀ for methane flux and methane production during decomposition of larch and birch wood (p < 0.05). For birch wood, Q₁₀ of methane consumption also significantly depended on a temperature interval (p < 0.05).

Calculated temperature response of CO₂ flux was 0.15–3.12 µg g⁻¹ h⁻¹ per °C and 0.004–7.64 µg g⁻¹ h⁻¹ per °C in the temperature interval +5 to +15 °C for birch and larch CWD, respectively. Under the temperatures from +15 to +25 °C, the response increased by 125–258% and 36–124% for birch and larch CWD, respectively (Figure 4). On the average this value was equal to 18.5–32.9% of the CO₂ flux measured at temperature +5 °C, and 13.4–39.1% of the flux at temperature +15 °C.

The average rate of methane flux decreased for the birch CWD at DC I and for the larch CWD at all decomposition classes when temperature increased from +5 to +15 °C, and for the birch CWD at DC II and DC III and for larch CWD decomposed by brown rot in case of the temperature increase from +15 to +25 °C. This reduction was 1.1% and 18.9% per °C in respect to the rate measured at +5 and +15 °C, respectively. In all other cases the rate of CH₄ flux increased with the rise of temperature and this increase was from 5% to 48% of the initial flux rate. The rate of methane production decreased only for DC III for larch and for DC II for birch CWD when the temperature increased from +5 to +15 °C. The average rate of this decline was 8.3% and 12.4% from the flux rate at temperature +5 °C. All other DC of larch and birch CWD showed respective increase of methane production with temperature rise from +5 to +15 °C (0–6.9% and 10.0–38.9%). With the temperature rising from +15 to +25 °C methane production increased by 12.8–46.3% and 24.3–53.8% per each °C for larch and birch CWD, respectively. Methane consumption decreased with temperature increasing only for DC II of birch CWD at the lower temperature interval. The rate of this process declined by 5% per °C. Larch wood at all decomposition classes and birch wood at DC I and DC III showed 2.1–25.7% increase of methane consumption per each °C increment when temperature increased from +5 to +15 °C, and 4.3–145% increment per each °C while temperature kept rising from +15 to +25 °C.

3.5. Effect of Water Content on CO₂ and CH₄ Fluxes

In our study, we analyzed CO₂ and CH₄ fluxes from decomposing CWD at their natural moisture. Water content in these samples varied from 14% to 176% of the mass of the dry wood for different stages of decomposition for larch CWD, and from 88% to 560% of the mass of the dry wood for birch CWD. Variations in CO₂ and CH₄ fluxes from these samples were connected with water content in CWD samples. However, the dependence of these fluxes on wood moisture was nonlinear (Figure 5). The respiration rate of larch wood was significantly positively related to the moisture of CWD samples when wood moisture was lower than 100% (R = 0.92 − 0.93 at temperature +5 °C and +15 °C, and R = 0.88 at +25 °C, p < 0.05). When moisture exceeded 136%, the CO₂ flux decreased, (Figure 5a). Birch CWD samples had higher water contents than the larch samples, and at DC II they also showed positive relationship between CO₂ and CH₄ fluxes and water content (R = 0.91 − 0.96, p < 0.05). Water content in decomposed birch wood at DC III exceeded 300%, and the increase of wood moisture from 320% to 500% showed the strong negative effect on the rate of CO₂ and CH₄ fluxes (Figure 5b).

Quadratic regression revealed that 45% to 53% of the CO₂ flux variance, and more than 50% of the variance of CH₄ consumption rate during decomposition of larch CWD, can be explained by the water content in the wood (p < 0.05). Methane production also depended on water content in the CWD samples; however, this relation was weaker than for CO₂ and methane consumption (R² = 0.33–0.47, p < 0.05) (Figure 5a).

The measured flux of CH₄ did not depend on water content in the CWD of both tree species. For birch CWD, the water content had a significant influence on the methane consumption rate at the temperature of +5 °C only (Figure 5b). In all the other cases the dependence on the water content was not significant.

3.6. Analysis of The Main Factors Affecting on CO₂ and CH₄ Fluxes

One-way ANOVA test showed that temperature alone was a significant factor for CO₂ flux from decomposing CWD, but it explained only 26% of respiration variance (R² = 0.26, p < 0.01). Temperature together with wood moisture explained 47% of variance for larch wood and 55% of the variance for birch wood ((R² = 0.47, p < 0.01) and R² = 0.55, p < 0.05, respectively). If we add to the model other influencing factors (tree species and decomposition class), it explains 62% of total variance of the respiration rate (R² = 0.62, p < 0.05) (Table S1). Separately for larch wood, these factors explained 51% of the variance of the respiration rate (R² = 0.51, p < 0.05), and 92% for birch wood respiration (R² = 0.92, p < 0.05) (Table S2).

Measured fluxes of CH₄ significantly differed for the two studied tree species (p < 0.05). Species and decomposition class were the main influential factors for the CH₄ flux from decomposed wood. All together they contributed about 24% to the total variance of this CH₄ flux (R² = 0.33, p < 0.05) (Table S1).

The calculated rate of CH₄ production significantly depended on the temperature (R² = 0.20, p < 0.01). ANCOVA analysis showed that water content, decomposition class and wood density together with temperature contributed about 44% to the total variance of this process. Temperature alone explained 15% and 45% of total variance of the rate of CH₄ production by decomposing larch and birch wood, respectively (p < 0.01) (Table S2). Altogether, temperature, decomposition class, water content, and wood density explained 44% of the variance of the methane production rate for larch wood (Table S2). Significant factors for the rate of CH₄ production during decomposition of birch wood were decomposition class, temperature and their interaction (p < 0.05). Altogether they explained almost 80% of the CH₄ production rate variance (R² = 0.79) (Table S2).

The rate of CH₄ consumption significantly depended on tree species, decomposition class, temperature, and their interaction (R² = 0.41, p < 0.01). Temperature alone explained only about 15% of total variance of the CH₄ consumption rate (R = 0.38, p < 0.01). (Table S1). Separately for larch CWD, temperature explained only about 10% of the CH₄ consumption flux variance (R² = 0.10, p < 0.05). Within all variables, the most influential for this flux was the water content. It contributed more than 20% (p < 0.01) to the total variance explanation (R² = 0.44, p< 0.05) (Table S2). For birch CWD, temperature alone explained 34% of variance of the methane consumption flux (p < 0.01). Together with other variables, the predictive capacity of the model reached 86% and the most significant factors of this model were temperature (39.9%, p < 0.01) (Table S2).

The rate of measured flux of CH₄ from decomposing wood had no correlation with CO₂ flux, while the calculated rate of CH₄ production was significantly (p < 0.05) correlated with the rate of CO₂ emission (R = 0.75 and 0.89 for larch and birch wood, respectively) (Figure 6). The rate of methane consumption also had a strong correlation with the rate of CO₂ flux both for larch and for birch CWD (R = 0.81, p < 0.05 and R = 0.96, p < 0.05, respectively) (Figure 6). CO₂ flux contributed 35% to the explained variance of the methane production rate (p < 0.01), and 51% to the total CH₄ consumption variance described by all factors (Table S3).

3.7. Effect of The Main Factors on Temperature Sensitivity (Q₁₀) and Temperature Response of CO₂ and CH₄ Fluxes

Temperature coefficient Q₁₀ for CO₂ and CH₄ fluxes did not differ significantly for the two studied tree species. However, ANOVA analysis showed that temperature sensitivity of CH₄ production significantly differed (p < 0.01) for the two studied temperature intervals. Temperature sensitivity of CO₂ flux significantly depended on moisture of CWD (6.7% of variance, p < 0.05). Within the lower temperature interval (from +5 to +15 °C) temperature sensitivity of CH₄ consumption significantly depended on interaction between tree species and decomposition class (p < 0.05).

Separately for larch CWD, the temperature interval explained 14% and 20.4% of Q₁₀ variance for CH₄ production and CH₄ flux (p < 0.05), respectively. Temperature sensitivity of CH₄ production depended also on decomposition class (15.4% of total variance of Q₁₀, p < 0.05). Temperature sensitivity of CO₂ flux significantly depended on water content (23.5% of variance, p < 0.05), decomposition class (13.5% of total variance, p < 0.01), and interaction between decomposition class and temperature interval (18.5% of total variance, p < 0.01).

Within the lower temperature interval (from +5 to +15 °C), the temperature sensitivity of CO₂ flux mainly depended on decomposition class (45.7% of total variance, p < 0.05), while under the higher temperature interval (from +15 to +25 °C), the main factor was moisture (48.3% of total variance, p < 0.01). Under the lower temperatures (from +5 to +15 °C), the Q₁₀ for methane consumption significantly depended on decomposition class, mass, and volume of the sample. These factors and covariates contributed 31.3%, 25.4%, and 18.3%, respectively, to the explanation of total variance (p < 0.05).

Temperature sensitivity of CO₂ and CH₄ fluxes from birch wood had no significant dependence on studied factors and covariates.

The temperature response of the CO₂ and CH₄ fluxes calculated from a linear model significantly (p < 0.05) differed for studied temperature intervals (from +5 to +15 °C and from +15 to +25 °C). Tree species influenced only temperature response of CO₂ and CH₄ production (p < 0.05), and temperature response of CH₄ flux and CH₄ consumption did not differ for the two studied species. Within the separate temperature intervals, the main factor influencing temperature response of CO₂ flux was the water content in the wood (42% and 30% to the total explained variance at the temperature intervals from +5 to +15 °C and from +15 to +25 °C, respectively, p < 0.05). Decomposition class of CWD additionally contributed 16% and 22% (p < 0.05).

Temperature response of measured CH₄ flux did not depend on any of studied parameters. In addition, temperature response of calculated CH₄ production did not show any significant difference in the temperature interval from +5 to +15 °C, but depended on decomposition classes at the interval between +15 and +25 °C (p < 0.05). Decomposition class as well significantly influenced the temperature response of CH₄ consumption when temperature increased from +5 to +15 °C (17% of variance, p < 0.05). The second significant factor was the water content (13.5% of variance, p < 0.05). Effect of moisture of CWD samples significantly increased in the temperature interval from +15 to +25 °C, and the water content explained more than 23% of variation in temperature response of the methane consumption (p < 0.05).

Separately for larch CWD, the water content explained 30% of temperature response of CO₂ flux at lower temperature interval (from +5 to +15 °C) and only 13% under the higher temperatures (from +15 to +25 °C). Temperature response of CH₄ fluxes mainly depended on decomposition class or density of wood in both temperature intervals. The main factor influencing temperature response of CO₂ and CH₄ fluxes from decomposing birch CWD was decomposition class, which explained from 46% to 83% of total variance in temperature response of this fluxes (p < 0.05).

4. Discussion

4.1. CO₂ and CH₄ Fluxes from Decomposing CWD

The rate of CO₂ emission from CWD samples was higher at DC I than at DC II for CWD from both tree species; however, for larch this difference was not significant at p < 0.05. The lower rate of respiration of logs at DC II can be a consequence of soluble sugars, starch, and non-structural carbohydrates depletion in decomposing wood [68]. Microorganisms consume these substrates at first. In addition, the wood-decaying fungi, which are the main decomposers of cellulose and lignin, do not invade widely into the solid wood at DC I. At DC III, the wood density decreased significantly and fungi and bacteria are able to penetrate the wood. Larger fungal and microbial biomass and resulting activity in wood at the advanced stages of decomposition lead to the higher respiration rates. Additionally, wood decomposition is accompanied by a relative nitrogen enrichment [69,70,71,72,73,74], which can promote decomposition rate of initially nitrogen-poor substrate [4,75,76].

Deciduous and coniferous wood differ in terms of morphology and chemical composition of the cell wall. Higher decomposition rates of deciduous wood correspond with their lower lignin concentration and higher nutrient concentrations in comparison with coniferous wood [4,76,77]. However, in our experiment, birch wood produced the significantly higher rate of CO₂ emission than larch only at the late decomposition stage under the highest temperature +25 °C. In other cases, CO₂ emissions from larch and birch CWD were comparable despite the higher nitrogen content and lower C to N ratio in the birch wood (Table 1). It could be a consequence of high water content in the birch wood that will be discussed below.

During incubation, we observed significant increasing of δ¹³C-CH₄ (up to +60.2–+74.0‰) and rising of the absolute concentration of ¹³CH₄ in the headspace of the experimental boxes. The similar ¹³C-CH₄ enrichment (from initial δ¹³C-CH₄ equal to −37.8 ± 0.6‰ to +84.6 ± 0.4‰) was observed in the incubation experiment on anaerobic oxidation of methane by Methylomirabilis oxyfera [78], and during aerobic microbial oxidation of CH₄ (from −22.0 to +76.4‰) in closed isotopic system [79]. These authors also showed that the progressive isotope enrichment was identical in all the cases independently on absolute biomass and methane content.

Light fraction of CH₄ (¹²CH₄) is easier oxidized in comparison with methane containing a heavier carbon isotope (¹³CH₄) [51,52,53], so that the latter can relatively increase in the experimental air volume, shifting δ¹³C-CH₄. Increase of the absolute concentration of ¹³CH₄ in the headspace of the experimental boxes indicated evident methane production during decomposition of studied CWD. An increase of the absolute concentration of ¹³CH₄ was observed in both cases when the concentration of ¹²CH₄ decreased due to methane oxidation processes and when content of ¹²CH₄ did not change or increase. We did not observe any changes of ¹³CH₄ concentration in the control boxes with and without water, so we concluded that production of this gas was related with decomposition of studied CWD samples.

We propose that methane produced inside wood could be oxidized by methanotrophs in wood pores. Because methanotrophs preferentially oxidize ¹²CH₄, then mainly ¹³CH₄ could reach the wood surface and go out to the headspace of the experimental box. As a result, headspace air in the box becomes enriched with ¹³CH₄ even if the content of ¹²CH₄ did not change (in the case if produced ¹²CH₄ was completely consumed inside wood) (e.g., birch CWD at DC II in Figure 3).

In some cases, when the rate of methane production is high or if there are not many methanotrophs in decomposed wood, some part of the produced ¹²CH₄ could also reach the wood surface and air around. In this case, we observed an increase of absolute concentrations both of ¹³CH₄ and ¹²CH₄ (e.g., larch CWD at all DCs and birch CWD of DC I in Figure 3). However, in this case, the isotope ratio in methane also can be enriched with ¹³C due to the fact that some part of produced ¹²CH₄ is consumed inside the wood anyway.

In other cases, high activity of methanotrophs can lead to the oxidation of ¹²CH₄ in the air of the headspace. In these cases, we observed an evident decrease of total methane content in the headspace of experimental boxes. However, at the same time, flux of ¹³CH₄ from decomposing wood, like that described above in the first case, could exist. This can result to a stronger enrichment with ¹³CH₄ than would be expected in the case of simple methane consumption from the air or increase of the absolute concentration of ¹³CH₄ (e.g., birch CWD at DC III in Figure 3).

We assessed the rate of methane production and methane consumption based on the assumption that evolved ¹³CH₄ reflects the rate of methane production due to its discrimination by methanotrophs.

Methane flux was higher at early decomposition stage in contrast to methane consumption that increased from DCI to DCIII for both tree species. Decreasing wood density and enrichment of CWD in nitrogen promote the increase of fungal and bacteria species richness at the advanced stages of decomposition [80]. In addition, as it was reported by Mäkipää with co-authors [81], the highest number of methanotrophs were found just in the late stage of decomposition. The methane production and methane consumption rates in most cases were higher for birch CWD than for larch, probably due to different fungal communities decomposing wood of these tree species [82].

Our incubation experiment demonstrated that, for both tree species, the stage of decomposition influenced the rate of CO₂ and CH₄ fluxes from decomposing wood and their temperature sensitivity and temperature response. Similar results were reported by Herrmann and Bauhus [83]. They studied temperature effect on CO₂ flux from decomposing CWD of three tree species and found that about 80% of the variation in respiration could be explained by species, wood density and moisture. In our study, wood density decreased from DC I to DC III (Table 1) and these classes showed different rates of CO₂ emission and different temperature sensitivity. Other studies also reported a significant effect of wood density on mineralization flux from CWD [55,56,60,61,84].

The strong positive correlation between CH₄ production rates and CO₂ emission indicated evident strong relationship of this flux with biological activity of decomposers in the dead wood. The significant strong relationship between CO₂ flux and CH₄ consumption is also a sign that oxidation of methane in decomposing wood is an active biological process. One of the products of CH₄ oxidation is CO₂; however, the rate of methane oxidation does not propose significant contribution of this flux to the total respiration. The calculated rate of CH₄ consumption is less than 0.01% of the total rate of CO₂ emission, so it did not significantly influence on the CO₂ efflux.

Methane flux from decomposing coarse woody debris was reported by numbers of authors [31,32,33,34,35,36]. They proposed that the main organisms responsible for this flux could be methanogenic bacteria and archaea that could survive in decomposing CWD due to the anaerobic conditions inside wood or that tree stems serve as a tube to release methane produced in the soil. In our study, CH₄ production was closely correlated to the rate of CO₂ emission, indicating a biological source of this flux and connection with activity of aerobic organisms. Evidence of aerobic methane production by wood-decaying fungi was reported earlier [39]. Our results showed that aerobic methane production is probably the main process contributing to the methane flux from decomposing coarse woody debris, because CH₄ fluxes had negative relation with oxygen deficit caused by high water content in decomposed wood.

Comparison with the rate of CO₂ emission from the soils of the studied region [85,86] indicated that CO₂ flux from the stock of decomposing CWD could comprise 2–5% of soil CO₂ efflux in forests undisturbed by fire during more than 100 years. Whereas, already one year after the fire, it increases significantly due to both the decrease in soil respiration and the increase in dead wood stock, reaching 82–139% of the soil flux. Methane emission from soil surface in these northern boreal larch forests was reported to be equal to 0.25–3.35 µg CH₄ m⁻² h⁻² [87]. Calculated possible CH₄ flux from existing CWD stock in old-growth and mature larch forests [19,85] can comprise 2–40% of the soil methane emission. In their research, Köster et al. [85] have found out that soil of all the studied postfire ecosystems (from 1 to >100 years after the fire) acted as a CH₄ sink. It means that CWD could be the only source of CH₄ in such ecosystems.

4.2. Effect of Water Content on CO₂ and CH₄ Fluxes

Water content in the wood was one of the most significant drivers of CO₂ and CH₄ fluxes in our experiment. Moisture availability is an important control of decomposer activity [88,89]. Water content affects the fungi inhabiting wood, influencing their growth and decay rates and the outcome of mycelial interactions [90,91,92]. Moisture is mentioned as a significant determinant of extracellular hydrolytic and oxidative enzyme activity [93,94,95]. The optimal and maximum values of wood moisture for decomposer activity are highly variable. For example, wood with the moisture content of 60–130% reported to decay rapidly; 90–120% may be the top limit for some decay fungi [96,97]; however, optimum values between 160–180% have been found for others [96]. Moisture content in the wood in our experiment varied significantly (from 14% to 560%), and for some samples it was lower than fiber saturation point (i.e., lower than 24–31% of dry weight) [96,98]. Consequently, these samples showed the extremely low rate of CO₂ and CH₄ fluxes. The dependence of CO₂ flux on the water content in the wood was unimodal—it peaked at intermediate water content and declined following increasing moisture, probably due to the rising of oxygen deficiency. It is interesting that methane flux and methane production also decreased under the high water content, despite that they should not depend on oxygen availability, but rather they should increase under the anaerobic conditions [37].

In our study, we analyzed CO₂ and CH₄ production from decomposing CWD at field moisture. Birch CWD samples had higher water content than larch wood, which could be attributed to the species features, because larch and birch CWD samples were taken at the one plot and at the same time, but they had quite different moisture due to the tree species peculiarities. Bulk density of the wood is considered to be a significant factor influencing water content, due to the increase of the water holding capacity of the wood with a decrease in density [99]. For our samples, water content strongly depended on bulk density only for birch wood, in contrast to larch wood that had the similar bulk density but significantly lower water content (Figure 7).

The peculiarity of birch logs lies in their strong and water-resistant bark, which remains on the log during the whole period of wood decomposition. This bark prevents wood inside the log from drying and provides a high level of wood moisture especially at the late stages of decomposition. It is possible that the community of birch wood decomposers (microorganisms and fungi) are well adapted to the partly anaerobic conditions inside overwetted wood, which may explain the high rate of respiration at a water content of more than 300% of dry weight. The rate of CO₂ emission from birch CWD decreased only at a moisture level exceeding 320%, while respiration flux from larch wood dropped already at significantly lower moisture (136%).

As it was found in the earlier studies, water content and temperature interactions could explain 61–80% of the observed variation in CO₂ flux from decomposing CWD [55,56,59,61,83,100,101]. In our study, interaction of temperature and moisture explained only 35% of respiration flux. However, moisture effect was stronger for larch wood, which had low water content and, probably, its decomposition was water limited. The most influential factors for birch wood respiration were temperature and decomposition class.

4.3. Temperature Sensitivity of CO₂ Flux

Low initial content of easily decomposable organic substances in wood [102] cannot suppose significant effect of this component on the rate of CO₂ and CH₄ fluxes during 20 days experiment especially for the late decomposition stages of CWD. It means that changes in the rate of CO₂ and CH₄ production at different temperatures can be attributed to the temperature effect on these fluxes. The Q₁₀ values calculated from measured CO₂ emission, under the temperature between +5 to +15° C in our study (2.87–3.51 for larch and 3.17–3.33 for birch), are comparable to the Q₁₀ values reported for coniferous (Korean pine) and deciduous (Amur linden) logs investigated by Wu et al. [60] (2.41–2.75 and 2.55–2.95, respectively). Yatskov et al. [77] also reported the stronger effect of temperature on the decomposition rate of CWD for birch (Betula pendula) as for coniferous tree species with more decay-resistant heartwood. For CWD in a European beech forest in Central Germany, Q₁₀ values of 2.2–2.7 for decay stages one to four were observed between temperatures of 0 and +20 °C [83]. For deciduous species in Michigan, USA [103], and for the conifer Picea mariana in Manitoba, Canada [55], Q₁₀ values of 2.20–2.57 were reported. The temperature sensitivity of CO₂ flux from CWD of pine, elm and oak logs in Southern Primorye, Russia, was 2.41, 1.89, and 2.28, respectively [104]. Based on the analysis of a global data set comprising the CWD decay rates and average annual temperatures, Mackensen et al. [105] showed the average Q₁₀ value equal to 2.53.

Non-linear temperature dependency of CWD respiration was also observed in many previous studies [56,60,61,83,100,103]. The exponential model was only applicable for temperatures below +22 °C. A decreasing temperature sensitivity with rising temperatures above +20 °C was observed in a laboratory incubation study by Wang et al. [55] and in field measurements of CWD respiration by Gough et al. [103]. The lower rate of Q₁₀ increase, nearly reaching a stable state, was shown as the temperatures increased from +18 to +36 °C. Consequently, Yoon et al. [58] proposed that the logistic model should be used to better describe the response of CWD respiration to temperature change. They also presumed that the decrease of temperature sensitivity of CWD respiration at high temperatures might have been caused by water loss under high temperature incubation.

In our study, the Q₁₀ value for CO₂ flux was smaller at the temperature interval from +15 to +25 °C as compared to that from +5 to +15 °C only for larch wood at DC I and II, and did not change for DC III. It means that for DC III the respiration rate did not reach a plateau at +25 °C, but continued to increase, despite that moisture of these samples varied widely, from very low 12.5–27.8% for wood decomposed by the white rot to higher values for samples decomposed by brown rot (29.8–136%).

Birch wood at DC I showed a decreasing Q₁₀ from lower to higher temperature intervals, the similar values of Q₁₀ for studied temperature intervals were observed for DC II. The late stage of decomposition (DC III) showed increasing Q₁₀ at higher temperatures. Such different patterns of temperature sensitivity changes probably reflect a succession of the decomposer community during decomposition of the wood. A similar increase of temperature sensitivity within the higher temperature interval was observed for larch wood at DC III decomposed by white rot. Temperature sensitivity of wood decomposed by brown rot decreased from lower to higher temperature intervals. Probably, decomposition of carbohydrates requires lower activation energy in comparison with lignin degradation [106]. Both tree species showed higher temperature sensitivity at lower temperatures at early decomposition stages. At the late decomposition stage, increasing lignin decomposition causes a shift of temperature sensitivity to the higher temperature interval. A similar shift of the respiration maxima at late decomposition stages to the higher temperature, in comparison with earlier decomposition classes, was found by Rinne-Garmston with co-authors [99].

4.4. Temperature Sensitivity and Temperature Response of Methane Fluxes

Methane exchange between decomposing CWD and the atmosphere is a result of two contrary directed processes: methane production and methane consumption. Both of these processes are dependent on temperature. However, the temperature sensitivity of these processes varied independently within two studied temperature intervals, producing differing ratios between the rates of these processes. Temperature coefficient Q₁₀ is one of the most common measure of temperature sensitivity. However, Q₁₀ is not the rate of change with respect to temperature, but rather the ratio between two rates at different temperatures [29]. Temperature coefficient Q₁₀ shows how much the rate of the process is altered with temperature increase by 10 °C. It is; thus, a relative measure that characterizes temperature sensitivity of the process but does not allow for recognition of the shift in ratio between contrary directed processes with temperature change. Additionally, coefficient Q₁₀ was not a good measure of temperature sensitivity of CH₄ fluxes, because, in some cases the rate of these fluxes was negative or equal to zero. In such cases, Q₁₀ inadequately reflected changes in the rate of CH₄ fluxes. It could be negative even in the case when the rate of the flux was changing from negative to positive or was equal to zero if flux decreased to zero at a higher temperature, or could not be calculated due to division by zero in case if the flux was equal to zero at a lower temperature. In addition, Q₁₀ can show which flux increases faster—CH₄ production or CH₄ consumption. However, Q₁₀ fails to show changes in ratio between the absolute rates of these two fluxes and does not allow to assess which CWD sample can emit more methane to the atmosphere [30].

Theoretically, temperature sensitivity is the rate of change in fluxes with respect to temperature [30]. Mathematically, it means that temperature sensitivity is the first derivative of the temperature dependence of studied process [29]. Calculated from the linear model, the temperature response showed how much, on average, the rate of the process changed with each degree of the temperature within the studied temperature interval.

To predict changes of CH₄ flux in respect to temperature change, the absolute values of the temperature response of methane fluxes are more useful than Q₁₀ values. The ratio between temperature responses of CH₄ production and methane consumption indicates whether methane flux will increase or decrease. If the temperature response of methane production is higher than that of methane consumption, then the rate of methane flux will increase with respect to temperature increasing (Figure 8).

For wood of both tree species at the early stage of decomposition (DC I), the temperature response of methane production was almost twice higher than that of methane consumption under conditions of temperature increase from +15 to +25 °C. This indicates that, in boreal forests, CWD of the early decay stage can serve as a source of methane flux to the atmosphere when air temperatures increase above +15 °C.

5. Conclusions

The results of the incubation experiment showed that CWD of larch and birch in the northern boreal forests can be a source of CH₄ flux to the atmosphere in addition to the CO₂ efflux. This is important because CH₄ is a more active greenhouse gas than CO₂.

According to the IPCC reports and prognoses [11], forest ecosystems in the northern boreal zone are expected to be subjected to a strong increasing temperature due to climate change. Our data suggest that the contribution of CWD to the greenhouse gas fluxes under expected climate changes depends on temperature and precipitation regimes. Increment of days with temperature above +15 °C can lead to the increase of CO₂ and CH₄ fluxes under conditions where the water content in the wood does not limit decomposition.

The temperature response of CWD respiration and methane production depends on the tree species. Another significant factor influencing CH₄ production and CH₄ consumption is the decomposition class. Disturbances or a general increase of tree death caused by climate change can enlarge the stock of fresh CWD. This can lead to the rise of CH₄ flux to the atmosphere, as CWD at the first stage of decomposition produces and emits the highest amount of CH₄.

We calculated the temperature response of methane fluxes based on the assumption of the linear relationship within the studied temperature intervals covering 10 °C. However, temperature dependence of CH₄ fluxes within these temperature ranges could be nonlinear. We think that additional studies are needed to assess the specificity of the temperature response of CH₄ fluxes within the studied temperature intervals and for a wider range of temperatures, which can be observed in the boreal region and is also expected in relation with prognoses of climate changes.

The results of our experiment showed a strong positive correlation between CH₄ production and CO₂ emission, indicating biological source and supporting findings on the aerobic character of the main process contributing to the CH₄ flux from decomposing coarse woody debris.